A human-computer interaction based path planning method for Mobile Robots in a complex environment

In applications of the navigation and control of unmanned ground vehicles in a cross-country environment and tele-driving a rover on the unknown lunar surface for scientific exploration, human-computer interactive path planning and planned path tracking is a significant way of teleoperation. In this paper, we use the method based on Rapidly-Exploring Random Trees (RRTs) to solve the robot path planning problems in a complex environment with crowded obstacles. Aiming at the problem of efficiency reduction that original RRTs planners have difficulties in automatically finding a resolution path in crowded regions of the robot´s configuration space, A novel artificial-guided RRTs (AG-RRTs) planner based on multiple RRTs will be introduced. The AG-RRTs planner improves the efficiency of connection and mergence of multiple RRTs by artificial-guided waypoints which mark the regions of narrow passages. In our experiments, the AG-RRTs planner yields much better performance in complex environments than original RRTs planners, which validates the effectiveness and practicality of this new method.